Smart Proposal Writing Assistant

Specific Problem Addressed

In proposal writing, teams often face dual challenges in data collection and content generation. Bidding information is scattered across multiple platforms, making manual search and organization time-consuming and prone to missing key details such as purchasers, winning amounts, or industry trends. Meanwhile, generating technical solutions, commercial responses, and service commitments lacks real-time data support, leading to inconsistent proposal quality and missed opportunities due to outdated information. For instance, quickly obtaining lists of bidding projects in a region, analyzing a company's bidding history, or parsing risks in tender documents requires efficient integration with bidding data APIs for improved accuracy.

Core Capabilities and How It Works

This skill leverages Shibotech's bidding data API, covering over 330 million real-time updated records nationwide, to address these issues through the following mechanisms:

• Data Query and Intelligent Processing: The skill automatically selects interfaces based on user queries, such as using searchProjectApi to search bidding projects, getZTBStructreDetail for structured details, or searchProjectContactApi for contract announcements. Interface mapping is managed via the references/api-mapping.md file for efficient calls, and references/national-keywords.json is used for keyword expansion and announcement type filtering.
• AI-Assisted Proposal Generation: Integrated with AI text processing interfaces like aiSearchSubmitPolling, it converts natural language requirements into executable search conditions and generates proposal content frameworks. For example, when a user mentions "implementation plan," the skill combines project timelines and team information to output suggestions.
• Key Steps: Users must first configure the environment variable BBIAO_API_KEY to enable API access. Once triggered, the skill parses query parameters (e.g., time range, region, amount), calls relevant APIs to fetch data, and organizes it into tables or bullet points, defaulting to the last 30 days of data for timeliness. For corporate profiles, it prioritizes querying companyProfileSummary to supplement customer and supplier analysis.
• Error Handling and Transparency: If API quota is insufficient (error code 0100590006), the skill prompts contacting sales for top-up and provides a fixed purchase page link; for authentication failures, it checks API Key validity. It does not automatically retry or handle payments, ensuring operational transparency, and outputs annotations like "not disclosed" to avoid fabricating data.

Applicable Boundaries and Considerations

While the skill effectively enhances the utilization of bidding data, the following points should be noted:

• Dependency on External Services: The skill relies entirely on Shibotech's API, so its stability is influenced by that service. The BBIAO_API_KEY must be configured correctly; otherwise, queries fail, and API quota limits require manual package management.
• Data Accuracy Limitations: The skill does not fabricate any announcements, amounts, or links; missing fields are annotated as "unknown." For broad terms, it combines industry signals, and low-relevance results suggest manual review, avoiding non-procurement content like recruitment or news.
• Scenario Suitability: Primarily designed for proposal writing, bidding queries, and corporate profile analysis, suitable for generating technical solutions and commercial responses, but contract-related issues need priority verification of contract announcement details, and near-building project opportunities require comprehensive judgment with bidding data.
• Operational Considerations: Output prioritizes user-provided keywords; default time, region, and announcement stages are auto-completed with explanations; error handling for insufficient balance only prompts purchase without automatically opening pages or polling.

Through this design, the skill provides structured data support while maintaining reliability and transparency, suitable for engineering teams to quickly access real-time information during proposal writing.